So, just how bad *ARE* transformers driving class B grids???

[N2DTS]

Why tie the grids together with a resistor?

Why not set the screen voltage and regulate it, and use a tetrode as a tetrode?

Is there some advantage to triode connecting tetrodes besides not having to provide screen voltaage?

Is there much reason to not use tubes that do AB1?

A pair of 4x150a's (4cx250b) will do 600 watts of audio in AB1, other tubes will do lower power levels, KT88, KT90's, and a quad of KT90's would likely do quite a bit of power in AB1.

4-125, 250 and 400's will do AB1, no?
Are they dirty?


Still, I do not notice any difference between the AB1 modulator and the 811A modulator, and looking at 1 and 2 tone tests on the air, I do not see much major difference. Not sure I see any difference at all.

What about broadcast rigs?
Did most run AB1 or something else?

[K1JJ]

In every case I've measured, drawing grid current in a linear tube configuration degraded the intermodulation distortion numbers.

This applies to grounded grid, class B push-pull audio and standard class AB grid driven RF linear amplifiers.  Unfortunately, without drawing recommended grid current, our power output is often reduced by 30% to 50% or more in some cases.

Running AB1 (no grid current) is how I run all my linear and modulator stages here.  The key is to start with a bigger tube(s) and run it AB1 to achieve the desired output for the cleanest results.  (With a stiff, electronically regulated screen supply, if used, drawing recommended current)

T

[steve_qix]

Good questions!

The major advantages of using a triode are a lower internal resistance (good for pushing audio through otherwise poor iron), and a larger plate swing (will get closer to 0V), so more modulation (positive peak) capability.

Old broadcast rigs vary with respect to the modulator.  The Gates BC1T I had used 833As in a class B modulator.  The RCA BTA1-R used 4-400s in AB1.

Both rigs sounded real good.

In AM broadcasting today everything is pretty much either a pulse width modulator or so-called digital modulation (like the Harris high power DX series of transmitters).  Haven't seen anything else out there, although there may be some.  There were some experiments using RF PWM, but have not seen a commercial transmitter using the method.....But, not being directly in the broadcast business any more, there could be some new methods that I've not seen yet.

[IN3IEX]

This tube:

http://tubedata.milbert.com/sheets/049/6/6KM6.pdf

That I use in my 70W AB1 modulator can give 200 mA anode current with only 10 V cathode anode voltage, 140V VG2 and -10V VG1.
This is much better than most classic power tubes. In fact it is a "modern" tube designed to provide a lot of current with almost no driving power.
See the curves, they are really good.
I use it pp with 600V anode supply voltage (1200V with audio peaks). Very satisfied.

Happy 2014 !!

Giorgio

[John K5PRO]

That is an interesting tube. With G3 connected to a pin, is it a suppressor grid?
G3 is energized to 30 VDC to reduce Snivet noise in television. What was that?

[w3jn]

A snivet is also known as Barkhausen oscillation. 

This guy says it better than I ever could:

" One of the phenomena that has appeared in the image reproduced in the television receiver is known as the snivet. Snivets are thin parallel vertical lines which may appear at one or more positions of the reproduced image. lt has been suggested that the production of such thin vertical lines is due to Barlthausen oscillations developed in the horizontal output tube.

"Barlrhausen oscillations may be produced in a tube when a positive electrode is arranged between two more negative electrodes. The electrons present in the tube are repelled by the more negative electrodes back toward the positive electrode. The oscillations caused by the alternate attraction and repelling of electrons occurs in a frequency range determined by the dimensions and spacing of the electrodes and the potentials applied thereto. Such oscillations may produce radio frequency radiation in the video carrier frequency range of the television receiver. T is radiation will then be detected and pass through such receiver in substantially the same manner as the video signal. The received radiation is then reproduced on the cathode ray screen in the form of the above-referred to thin vertical lines. "

Barkhausen oscillation can be evident in high-gain beam tubes even in a tube tester.

[N2DTS]

A snivet is also known as Barkhausen oscillation. 


I have that.
We have an English Setter, and any time the mail man or UPS guy comes around, the dog starts that oscillation.
I suppose the mic has picked it up sometimes and its gotten into my tubes.

[WBear2GCR]

Steve,

So the nonlinearity you show in the scope image occurs only when the modulator grids are driven positive and draw grid current?

First thing I would try is to load that transformer with a dummy load and test. I'd want to be sure my "50 watt" amp was not running out of current. The transformer has a reactive component which could be a factor.

The idea already mentioned about the impedance of the grid changing, dropping radically when current is drawn is another factor. In effect you now have a massive mis-match at the output side of the transformer (the former secondary). That changes the impedance seen by the primary downward, and your "50 watt" amp may not like that at all.

The dummy load would have to reflect that, so you could test low Z and high Z.

Try a Crown Macrotech or something like that, odds are the triangle stays solid.

                            _-_-

[steve_qix]

Hi Bear,

Look back at the very first post, and there is a picture of the oscilloscope waveforms showing the input to the transformer (a perfect triangle wave), and the output (distortion).

I also thought the 50 watt amp was running out of soup, but in fact all was well at the output of the amp!!

[WD5JKO]

Steve,

    Looking at that original waveform, it appears to load down as a result of the 1625 triode connected grid current. That transformer used might have several hundred ohms DC resistance on each side of what was the KT88 primary. The winding technique used might have an inner and an outer layer such that the winding DC resistance would be less on the inner winding, and more on the outer winding when each has the same number of turns. If this is the case, then one 1625 grid should show less effect from the grid current then the other. If each grid waveform is close to the same, then the transformer coupling is lousy from high leakage inductance.

  I'd wager that same transformer turned around (step-down) from a high voltage SS driver (FET's running at about 400V B+ in push pull with NFB) would do much better. Here we use the 4 ohm tap as a CT, where the common, and 16 ohm wires feed to the 1625 grids. The turns ratio might be a little steep though, so I don't know if the peak grid to grid voltage could be realized this way unless the SS driver B+ is boosted up further.

  Back in the day, they never used a step up transformer to feed class B grids, mostly due to the adverse effect on the driver that had a much higher source impedance from the tubes used. The best way around this was a step down impedance transformation.

Jim
Wd5JKO

[steve_qix]

Hi Jim - absolutely agree !!

I just put the note here because a lot of folks actually use reverse-connected audio output transformers to drive class B grids - hear people talking about it on the air.  And of course I tried it myself with the documented results.

It's probably time to lay out the PC board for the "universal class B driver".  It's designed and I built a prototype, but haven't had the time to lay out a board.  The universal driver is a direct coupled MOSFET affair that will do around 500V P-P into a varying impedance.  It's fully overload and arc-back protected.

My goal is to make it sufficiently small so as to be able to fit under the chassis of a Ranger, Valiant, etc., but still be able to drive tubes like 833As and the like.

[kb3ouk]

I wonder what would happen if the class B stage was grounded grid, cathode driven? I've seen schematics for single ended audio amps that were grounded grid, and push-pull RF amps that were grounded grid, so I would assume that a push-pull GG audip amp would be possible. A tube with an indirectly heated cathode would be easy to do, I'm not sure how you would drive a tube with a directly heated cathode though, my guess would be seperate filament transformers with the leads of the driver transformer's secondary feeding the center taps of the tubes. Driver transformer would have to have a center tap on the secondary side, or else the cathodes would be floating.

[WU2D]

I have always used iron but a conventional phase splitter into couple of 6N7's s as White cathode followers should work to drive the grids. Sort of a weight vs complexity issue.

[W2VW]

I wonder what would happen if the class B stage was grounded grid, cathode driven? I've seen schematics for single ended audio amps that were grounded grid, and push-pull RF amps that were grounded grid, so I would assume that a push-pull GG audip amp would be possible. A tube with an indirectly heated cathode would be easy to do, I'm not sure how you would drive a tube with a directly heated cathode though, my guess would be seperate filament transformers with the leads of the driver transformer's secondary feeding the center taps of the tubes. Driver transformer would have to have a center tap on the secondary side, or else the cathodes would be floating.

It's do-able just as you describe.

[WBear2GCR]

You may still have a problem with an indirectly heated tube run with GG (cathode drive) because you will be swinging the cathode and possibly exceed the filament to cathode max voltage.

The trick in the case of the DHT would be to swing a floating filament supply (or let it float) and would do the same job for the indirectly heated cathode, I would expect.

Steve, I am still puzzled by the results you are getting and would really like to see that OT loaded by some resistors of different value, until you get the same looking result. The grid ought to require no more than a watt or so at most, and I don't think this explains what you are seeing exactly.

Noted is the input waveform looking ok.

                 _-_-